More comprehensive statistical models will be developed for the genetic analysis (segregation and linkage) of breast cancer susceptibility in human families and pedigrees. These models will allow for environmental risk factors, variable age of onset, and a mixed monogenic-polygenic mode of inheritance. Algorithms will be developed to make the accurate calculation of the likelihood of multivariate pedigree data feasible under these models, to allow the testing of specific genetic and environmental hypotheses. Based on these models and algorithms, computer programs will be written in standard ANSI X3.9 - 1978 (FORTRAN 77) and documented. Computer simulation will be used to investigate robustness and power. These methods of analysis will eventually be applied to pedigree data that are currently being collected, with a view to confirming or refuting the existence of a major gene for breast cancer susceptibility.